Study On The Technology For Classifying Damage Level Of Plateau Pika (Ochotona Curzoniae) To Alpine Rangeland Based On UAV Remote Sensing

Plateau pika is a unique small mammal in the Qinghai Tibet Plateau.It is the key stone species in alpine rangeland ecosystem because it plays an important role in the energy flow and material circulation.However,when the population of plateau pika is too high,the pika has the negative impact to the production and ecological function of rangeland ecosystem due to their behaviors of foraging and digging holes.Therefore,it is necessary to monitor the population distribution,density of plateau pika and evaluate the damage level of the pika to alpine rangeland.The traditional monitoring methods of plateau pikas'damage are mainly rely on manual investigation,which has the disadvantages in term of efficiency and accuracy.Recently,UAV(Unmanned Aerial Vehicle)technology has been wildly applied in the ecological research because of the advantages of convenience,rapidity and efficiency.In this study,we used UAV technology combining with investigation in field in Maqu County of Gansu Province to monitor and evaluate the damage level of plateau pika.The purpose of the study is to develop the technology using UAV to classify the damage level of plateau pika to rangeland.The study has three components.No.1,we had determined the optimal flighting parameter in term of aerial height,aerial time,interpretation mark,etc.for monitoring the pikas'damage.No.2,we had developed the technologies of UAV photo interpretation for monitoring the number of holes,aboveground biomass,grassland coverage.We also tested the accuracy of the interpretation technology in filed.No.3,based on the information of UAV photo and edible forages proportion in filed,we developed the comprehensive index to classify the damage level of plateau pika in the study area.The main results are as below:1.The optimal aerial height was 20 m,the optimal month of photo shooting in whole year was at returning green stage,the optimal time of photo shooting in a day was in noon.2.Using Python,we developed a software to interpretate the pikas'hole number,and plant cover from UAV photos.The accuracy of photo interpretati on was 95%and 93%,respectively,which were verified in field.Using ArcGI-S software,we selected the VDVI(Visible-Band Difference Vegetation Index)was the optimal index for estimating the aboveground biomass and developed a model(y=471.08x~2+25.47x+9.9703),which the accuracy was 86.54%.3.Based on the pikas'hole number,plant cover,aboveground biomass and the proportion of edible forage,we developed an integrated index,which was plateau pika damage rangeland index(DRI),and classified the pikas'damage rangeland in the study area into five levels,which involves Level I(no damage),Level II(light damage),Level III(moderate damage),IV(severe damage)and V(extremely damage).